Doctors’ AI scribes get names of drugs and diagnoses wrong, NHS watchdog warns
who could have fucking predicted this at length the second it was announced
Discussion
Doctors’ AI scribes get names of drugs and diagnoses wrong, NHS watchdog warns
who could have fucking predicted this at length the second it was announced
@davidgerard My physician, at a university medical center, said the school had them use AI transcription for a year. Everyone hated it, for the reasons you cite and others, people quickly stopped using the transcripts, and that was that.
@davidgerard Probably don't need a lot more reason to remove abusive power.
Don't worry, when the NHS rolled this out they required GPs to accept legal liability for inaccuracies, so I'm sure overworked GPs are reading the transcripts in detail! Oh, they're not? Well, at the very least, the companies selling this have a liability shield to hide behind, and that's the important thing.
@david_chisnall the doctors frequently fucking love it, and i can fully understand why. one of these that *worked* would be a godsend. but uh.
@davidgerard I think it is a variant of the Gell-Mann Amnesia effect. You forget that all the mistakes you see AI make apply to /everything/ it does. Although management may just be too stupid to have recognised the mistakes....
@davidgerard Haven't you heard?
AI doesn't need organic life.
@davidgerard Who signed off untested AI for any health staff?
It is fundamentally not trustworthy - not least it has no feelings and hence no sense of caring.
@davidgerard "More than half the 1,003 UK GPs in her survey last year believed that their ambient AI records were more accurate than those produced themselves, Blease said."
People believe a lot of things. More than half the voters in the USA thought Trump should get a second term. Just because a lot of people believe something doesn't make it true...
@davidgerard well who could have seen this one coming.
@capnthommo if only a few thousand more people had pointed out the obvious
@davidgerard yes & in some places do not have universial #healthcare & it should be their #right
@davidgerard I have a friend in the US who was asked by his doctor about marijuana usage. My friend replied truthfully that he does not use it. The AI scribe recorded that he uses it habitually.
@davidgerard
How every conversation with GPs has gone lately.
"I would like to use an AI app to take the notes of our conversation today. Are you okay with that?"
"On two conditions. First, can you send me a copy of the notes?"
"Certainly. And the second?"
"If I point out inaccuracies, can you send me a copy of the corrected file?"
"...I am not sure how that would work. I'll have to talk to the practice manager."
<...crickets from practice manager>
@Hedgewizard i just told mine "no thank you, i work in AI so i know *precisely* why I can't trust it"
(then i tried hard not to be a PITA otherwise)
Why are you mad? It's helpfully shortening the long boring report by deleting words that completely change the meaning!